This is a place to collect some of our ideas about "Symbol Systems", Chapter 2 in A Clark's (2001) Mindware (pp. 28-42).
I found the following link to a 10-minute video presentation of the SOAR system (Clark, pp. 31ff.) to be helpful in making the discussion in the text more concrete. Please take 10 minutes and watch it.
[michelle] Interesting...in the video, it looked like there would have been two (or more?) ways for SOAR to accomplish the given task, but only one was discussed. I think that the path SOAR chose had the shortest number of moves, but was it just as likely that it would have chosen the other path if it happened to search in a different order? Or does it have a way of comparing possible solutions to find the one with the lowest "cost"?
[danny] Right Michelle, I was wondering the same. Seems that this a modified depth first search algorithm that while is complete (will find a solution given a non-infinite branch depth), is not necessarily optimal since there was no talk of an evaluation heuristic used to decide with selection might give the shortest solution.
Please also be aware that the usage of the word "chunking" in the video and in Clark (pp. 31-33) is not the same as Miller's (1956). So much for standardizing our "technical" terminology . I suppose the word has been around long enough in cognitive science that people are now feeling free to use it in more or less "metaphorical" senses. This is, to be sure, the way language evolves, but it can be confusing to someone who is not well-versed in the field (hence my caveat here).
What is the "cognitive band"? (Rosenbloom, cited in Clark p. 33) Answer: It's the band between the "biological band" and the "rational band". I found the following picture helpful:
I found the following link to a 10-minute video presentation of the SOAR system (Clark, pp. 31ff.) to be helpful in making the discussion in the text more concrete. Please take 10 minutes and watch it.
[michelle] Interesting...in the video, it looked like there would have been two (or more?) ways for SOAR to accomplish the given task, but only one was discussed. I think that the path SOAR chose had the shortest number of moves, but was it just as likely that it would have chosen the other path if it happened to search in a different order? Or does it have a way of comparing possible solutions to find the one with the lowest "cost"?
[danny] Right Michelle, I was wondering the same. Seems that this a modified depth first search algorithm that while is complete (will find a solution given a non-infinite branch depth), is not necessarily optimal since there was no talk of an evaluation heuristic used to decide with selection might give the shortest solution.
Please also be aware that the usage of the word "chunking" in the video and in Clark (pp. 31-33) is not the same as Miller's (1956). So much for standardizing our "technical" terminology
What is the "cognitive band"? (Rosenbloom, cited in Clark p. 33) Answer: It's the band between the "biological band" and the "rational band". I found the following picture helpful: